The tautology p - q - p is not a theorem of the various relevance logics (see Anderson and Belnap ) because q is not considered to be relevant in the derivation of final p. We can take this lack of relevance to mean simply that p-q-p could have been proved without q and its -, i.e., p-p. By the same criterion we could say that in ((p-p) -q) -q p-p is not relevant. In general we will say that any theorem A of an implicational logic is strongly relevant if there is no subpart B ! which can be removed from A, leaving the rest still a theorem of the same logic. Such a subpart B - is said to be superfluous.
Bunder, M. W. (1994). The strong relevance logics. Bulletin of the Section of Logic, 23 (1), 12-17.